The present invention generally relates to communications systems, and specifically to an improved system and method of providing guaranteed bandwidth on demand for an end user and/or enterprise.
Internet protocol (IP) networks were designed to handle any traffic, on any port, at any time. The goal was to utilize as many computing platforms as were available across a consortium of universities, governments and industries willing to share information (Reference IETF RFC 791 Internet Protocol Darpa Internet Program Protocol Specification, September 1981).
With these goals in mind, each of the computing platforms, or routers, were originally designed to be ad-hoc in nature. That is, they broadcast on each of their ports, the routing and cost to send a packet to itself. Each manufacturer of these routers defined their own concept of cost and its associated value. As a result of IP's original design goals, the path that a packet takes from origin to destination is completely unpredictable. In the example in FIG. 1, a user is attempting to send IP video packets streaming from a source 100 to a destination 102, but the originator can not predict nor control how those packets will be transported across an IP network 104, nor can the originator even assume that all the packets streamed through the network 104 will take the same path to reach the destination 102. An IP router can not plan how a packet (or stream of packets) will reach its destination, nor can routers plan how many other routers will transmit the packet. It takes, on average, anywhere from 10 to 20 or more routers to send a packet across the internet today.
Now referring to FIG. 2, every router stops each incoming packet to determine whether it is allowed, its class of service, how to route it, and then, because they are processing so many unpredictable packet sizes/rates, they must queue the packets at both the ingress 200 and egress 202 ports, and possibly even at the internal switching matrix 204. A typical IP router architecture includes packet switching matrices 204, intelligent routing processors 206, and large memory queues at the ingress 200 (incoming) and egress 202 (outgoing) ports, as well as at a centralized interconnect level to move packets from one ingress port card to a different egress port card. With so much queuing and processing on each packet, packets may be lost or delayed beyond video services quality tolerance.
The services that may be delivered on broadband are many, ranging from real-time critical applications for communication purposes: video calling, multi-player gaming, telemedicine, television studio broadcast interviews, and high-definition news multicasting to name a few. These examples and a few others are listed in FIG. 3. These real time critical applications are very sensitive to any delay and for any that may include video or gaming frames, very sensitive to any variance in the delay. Applications which include video are also sensitive to any packets (or frames) which may be lost in the transmission (0.0001% packet loss is the preferred quality for video transmission).
Multi-Protocol Label Switching (MPLS) was developed to overcome some of the traffic engineering constraints of the IP protocols. MPLS allows operators to engineer a core network that aggregates traffic from IP, ATM, Frame Relay or even time-division voice domains, across a common packet core network. MPLS network operators can pre-define label switch paths, and ensure that virtual private network traffic is delivered on specific routes to achieve guaranteed quality of service levels (See IETF RFC 2702, Requirements for Traffic Engineering over MPLS).
MPLS standards have expanded to include point-to-multipoint multicasting (Reference IETF 4461: Signaling Requirements for Point-to-Multipoint Traffic-Engineered MPLS Label Switched Paths (LSPs)), and resource reservation protocols (Reference IETF RFC 3209, RSVP-TE: Extensions to RSVP for LSP Tunnels and RFC 4420) that dynamically utilize bandwidth across the core thus enabling less expensive transport for video broadcast traffic. The multicasting protocol enables construction of a distribution tree that replicates packets only at the branch points, rather than from the origination point. Now referring to FIG. 4, a stream of packets can begin at a single source point in the IP domain, and traverse across an MPLS packet network starting at a point 400, following a controlled path to a specific router at point 402, bypassing any un-necessary MPLS routers like point 404. The Originating MPLS Router can utilize the point-to-multipoint multicasting capabilities of MPLS, to instruct MPLS Router 402 to multicast the traffic to another user connected to MPLS Router 406. MPLS also expanded to include a Fast-Reroute method, which allows for a 50 millisecond route recovery in the event of a link failure, comparable to that of optical SONET networks. These attributes make MPLS the technology of choice for core network video transport today.
However, MPLS does not readily extend to the customer premises locations, as its focus has been on core packet transport aggregation, enabling controlled routing and quality of assurance through the packet transport. Also, MPLS was developed around the concept of delivering enterprise virtual private networking; thus much of the protocols and methods of packet quality assurance in MPLS require the utilization of a virtual Local Area Network (LAN).
Although IP Multimedia Subsystem (IMS) standard protocols evolved to try to address handling real-time multimedia streams across the IP packet domain, these standards have largely focused on enabling the streaming services as an overlay solution across existing IP network domains, without addressing any changes to the IP or MPLS routing architectures. Quality assurance requires managing the services end to end, from customer access point to access point. In addition, IMS standards were intended to be access agnostic, so the customer premises access point standards have been separately handled by various wireless (CDMA, GSM, UMTS, WiFi, WiMax, etc.) and wireline (Cable, DSL and Fiber, etc.) access standards.
Recently, focus for broadband applications has moved away from IMS to an evolution of these protocols within the 3GPP organization called TISPAN (Telecommunications & Internet Converged Services and Protocols for Advanced Networking) TISPAN intends to include methods for handling resource allocation and quality assurance, but again does not address the elements that sit within the customer premises to network access domain, leaving those up to the other standards bodies governing the various access types.
For the current broadband services deployments taking place, broadband network operators are utilizing mechanisms like the IEEE 802.1p bit marking to differentiate the service classes, and route traffic accordingly. Now referring to FIG. 5, the current services, comprising legacy public switched voice 500, video 502 and best-effort internet 504 access are served by existing network components, interconnected to the access networks via ATM, IP or IP/MPLS routers 506 and/or optical multiplexing solutions 508. Consumers and/or enterprises 510 connect via an access network 512, broadband or narrowband, to the services domain through access network equipment such as DSL Access Multiplexors (DSLAMs), Fiber Optic Access (such as Optical Line Terminals-OLTs) and various other access technologies. Services are delivered with assurance by interconnecting to the consumers via the broadband access network utilizing technologies such as IEEE 802.1 p bit defined service types. There are 8 p bits to differentiate service type—thus only 8 service classes. This is insufficient to cover a multitude of service offerings that may all require high quality broadband connections.
Today, the only quality video transport with assurance that operators can use are dedicated line, virtual private networking services. Each new service that requires a high quality packet transport requires a separate virtual private network. This does not allow for dynamic bandwidth allocation and utilization—thus it does not economically scale across multiple services or across multiple users. An example of is illustrated in FIG. 6.
Video transmission requires compression in order to effectively utilize the available broadband bandwidth across packet domains. Currently there are numerous different methods for encoding the video, some standardized and some are proprietary. Many existing video communication solutions today utilize proprietary mechanisms, which are incompatible across multi-vendor and access domains. Additionally, the video compression methods vary greatly in the bandwidth they require to transport the video in real-time—some solutions are as low as 64 kbps up to 300 Mbps. The bandwidth required can vary based on the codec type and the quality type compressed within the codec type. For example, MPEG-4 (Motion Picture Experts Group-4) defines methods to combine and encode video with sound and text, including the encoding of Standard Definition and High Definition.
Therefore, what is needed is an improved method and system of delivering guaranteed high bandwidth applications to an end user and/or enterprise end to end.